On Fri, Jan 25, 2008 at 01:01:54PM -0800, David Bustos wrote: > But you still intend to control, just through the administrator's > interfaces rather than the restarter interfaces, right? I'd rather > augment the restarter interfaces than have your service use the > administrator's interfaces programmatically. Do you see what I mean?
I don't. We have Committed SMF admin interfaces. They aren't just aren't APIs. But why shouldn't we have admin APIs? I think that torturing SMF event consumer/admin wannabes into having to be restarters, even part-time restarters, complicates things a lot. Besides, it's a serious and, to my eyes, unnecessary constraint: Q: How many services can be the restarter for any given service instance at any given time? A: Just one. I can see how a mechanism for excluding multiple copies of the same event consumer could matter, but then again, we already have that mechanism. (BTW, a monitor sub-command for svcs(1), much like the monitor sub- command of route(1M), strikes me as useful too.) Nico --
